Professional Documents
Culture Documents
1102.eps
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means without express written permission from Industrial Fiber Optics.
TABLE OF CONTENTS
ACTIVITIES COMPONENT IDENTIFICATION ..................................................... MAKING A LIGHT GUIDE................................................................. FIBER O PTIC C ABLE TRANSMISSION ................................................. CONNECTORS AND SPLICES ............................................................ 1 2 5 8
INDEX M ATCHING ......................................................................... 11 FIBER TERMINATIONS .................................................................... 14 SPEED OF OPTO -ELECTRONIC DEVICES ............................................. 17 FIBER O PTIC TRANSMITTERS ........................................................... 20 RECEIVER AMPLIFIER DESIGN .......................................................... 24
-i-
- ii -
- 1-
Table 2.2 Empirical data for 10 cm light guide with water core. Source Red LED OFF 0* LED ON .61 mA
Table 2.3 Empirical data for 10 cm light guide with90-degree bend. Source Red * LED OFF 0* LED ON .46 mA
This current should measure very close to zero, or less then 1 A .If the current is not this low, dim the room lights or cover the light guide and associated electronics with a dark cloth to obtain best results.
phototransistor illumination with water core = .61 mA/ (50 mA/mW) = 12.2 W phototransistor illumination with no water in core = .02 mA/(50mA/mW) = 0.4 W
- 2-
Does the light guide couple more or less light onto the phototransistor with water in the core? Why? More light .The water has a higher refractive index than air, so less light is lost at the boundary with the vinyl tubing. As a result more light travels down the light guide from the LED to the phototransistor.
Did the 90-degree bend significantly change the amount of light hitting the phototransistor? Why or Why not?
The 90-degree bend did reduce the amount of light hitting the phototransistor. The bend caused some of the light rays traveling down the light guide to fall inside the critical angle for total internal reflection. These rays escaped the light guide, reducing the amount of light falling on the phototransistor .
Calculate the critical angle of the light guide when water is used as the core. Assume the refractive index of water is 1.33.
c = sin 1
1 1.33 = 48.8
- 3-
Do you know of any other liquids that may trap more light inside the vinyl tubing than the water used in this experiment? To trap more light inside the vinyl light guide one should choose a liquid with a higher refractive index than that of water. The higher the refractive index, the better the liquid will trap light (without considering the effects of attenuation). To find more information about liquids with a high refractive index, one of the best areas to look is the reference section of a library. A common book found in this section is the " Handbook of Physics and Chemistry." Examples of liquids with a refractive index greater than 1.33 include Pentachloroethane (n=1.501), Xylidine (n=1.557) and Quinoline (n=1.622).
c1
R1
c2
R1' R2
c2
c2
HOMEWORK PROJECT
Following are some typical fiber core/cladding outer diameters: Single mode glass fiber: 9/125 m or 8/125 m Multimode glass fiber: 50/125 m 62.5/125 m 100/140 m Multimode plastic: 230/250 m 480/500 m 730/750 m 980/1000 m (fiber used in the kit) 1980/2000 m
- 4-
i3 meter
8.2 mA .117 mA .416 mA
i1 meter
10.2 mA .143 mA 3.53 mA
Are the attenuation coefficient values the same for the different LEDs? Why or why not? No. The optical fiber has an acrylic core in which light transmission varies for different wavelengths. The attenuation coefficients that students will calculate from their empirical data should show the coefficients for the red and the green LEDs to be very close to each other. The attenuation coefficient for the infrared LED should be much higher than the coefficient for the red or green LEDs.
- 5-
Calculate the launch power for each LED using Table 3.3 . (Substitute the phototransistor current for power in this equation because the phototransistor current is linearly proportional to the optical power.) Table 3.3 Calculation of launch power (equivalent current) for each LED. LED Red Green Infrared
P 1 meter
10.2 mA .143 mA 3.53 mA
e (1)
1.119 1.106 2.913
(1) Po = P 1 meter e
Calculate the phototransistor current produced by the light from the three LEDs traveling down a 5meter plastic optical fiber. Use Table 3.4 as a guide. (Obtain a from Table 3.2 and Po from Table 3.3.) Table 3.4 Calculation of phototransistor current for a 5-meter fiber length. LED Red Green Infrared
Po
11.41 mA .158 mA 10.28 mA
l
5 5 5
e ( 5)
.571 .604 .0048
P5 meter = Po e ( 5)
6.515 mA 0.095 mA 0.049 mA
Calculate the phototransistor current produced by the light from the three LEDs having traveled down a 10-meter plastic optical fiber. Use Table 3.5 as a guide. (Obtain a from Table 3.2 and Po from Table 3.3.) Table 3.5 Calculation of phototransistor current for a 10-meter fiber length. LED Red Green Infrared
Po
11.41 mA .158 mA 10.28 mA
l
10 10 10
e (10 )
.326 .364 0.000023
(10 ) P 10 meter = P 0 e
- 6-
Plot measured phototransistor currents for the 1-meter and 3-meter lengths, and the calculated phototransistor currents for the 5- and 10-meter fiber lengths for all three LEDs in Figure 3.2.
Versus
Fi gu re 3.2 P hototr an sistor c ur ren t c rea ted a t th e en d of var iou s fi ber len gth s for th ree di fferen t LE D ll ig ht sour ces. Although this question did not ask students to include the LED launch power (Po) on the chart, it is included on the one above. The purpose of the graph was to illustrate that the infrared LED initially produced the most phototransistor current and the green LED the lowest. Then for any distance greater than 5 meters the red LED produced the most phototransistor current. The most commonly used wavelength for long distance transmission in plastic fiber is the 650/660 nm window. This wavelength band offers the lowest material attenuation and readily available LEDs with good optical output power. The graph above verifies that the 650/660 nm wavelength band is the best choice to maximize the working distance with plastic fiber for the three light sources tested.
HOMEWORK PROJECT
The color red has a wavelength from 630 to 670 nm and the color green from 490 to 540 nm.
mA
- 7-
isimplex
3.5 mA 0.22 mA 0.20 mA 0.17 mA 0.19 mA
Table 4.2 Measured phototransistor current with 2-meter and 1-meter optical fibers spliced together. LED Red Infrared
isplice
3.9 mA 0.25 mA
- 8-
What happens to the measured phototransis tor current when the simplex assembly and 1-meter fiber are rotated to different positions within the simplex receptacle? Describe below the physical conditions that are occurring and why. (Drawing a picture might be helpful.) The phototransistor current changes. As the simplex assembly and fiber are rotated in 90 steps, changes occur in the lateral and angular alignment which cause variations in the amount of light coupled from one fiber to another. These physical variations cause changes in the amount of coupled light which travels down the optical fiber and is absorbed by the phototransistor. Table 4.3 Comparison of transmission characteristics of a continuous 3meter fiber optic cable to those of a 3-meter fiber length with fiber optic connector installed. LED Red Infrared #1 #2 #3 #4 Rotation
iActivity
8.2 mA
III
isimplex
3.5 mA 0.22 mA 0.20 mA 0.17 mA 0.19 mA
% Transmission
Table 4.4 Comparison of the transmission characteristics of a continuous 3-meter fiber optic cable and those of a length with a splice in it. LED Red Infrared
iActivity
8.2 mA
III
isplice
3.9 mA 0.25 mA
% Transmission
47.6 60.1
.416 mA
Is the transmission greater for the 3-meter fiber with the splice installed or with the simplex receptacle? Is this what you expected? Why or why not? The transmission is greater, or there is less loss, for a splice. Yes. Because the physical tolerances are smaller for the splice, which tends to align fibers more accurately and allows light to be better coupled from one fiber to another.
- 9-
In your own words, state at least two advantages and disadvantages of fiber connectors versus fiber splices. List at least two for each. The advantages and disadvantages of fiber splices and connectors can be described in many ways. Below is a comparison of main features of splices and connectors from which you can interpret students' answers. Connectors Non-permanent Factory installable on fiber cables Easy reconfiguration Simple to use Field installable Permanent Easier to obtain low loss in field Lower attenuation Spliced fibers can fit inside conduit Less expensive per interconnect Stronger junction Some are hermetically sealed Splices
HOMEWORK PROJECT
We suggest that you review students' company selections so that their information requests arrive in time to complete A CTIVITY IX, and so you can ensure the selections they have made are within the scope of your class.
- 10 -
isplice
3.9 mA 0.25 mA
iindex matched
6.0 mA 0.36 mA
isplice
3.9 mA 0.25 mA
iindex matched
6.0 mA 0.36 mA
% Improvement
53.8 44.0
Explain the change in transmission indicated by Column 4 in Table 5.2 . Glycerin improved the transmission because its index of refraction was very close to that of the fiber core. This reduced the Fresnel reflections, compensated for less-than-perfect fiber ends, and reduced light spreading at the splice. The glycerin in this activity acted as a refractive index-matching compound to demonstrate the procedure commonly called "indexmatching" in the fiber optics industry. The students' most common answer should be "to reduce Fresnel reflections." Explain in your own words what "index-matching" means. An index match compound is liquid or gel with a refractive index close to that of fiber core materials. It eliminates the fiber core/air boundary conditions in splicing and is occasionally used in connectors. Its main purpose is to reduce Fresnel reflections (although it sometimes reduces other losses as well).
- 11 -
Index-matching is sometimes used as a verb and refers to the procedure of using an index-matching compound to minimize the loss between the junction of two fibers. Calculate the magnitude of a Fresnel reflection using the equation at the introduction to this activity. Assume the refractive index of Medium #1, the fiber core, to be 1.49 and Medium #2, air, to be 1.0. The ratio of the reflected power to the incident power is:
R = 2 1 2 + 1
2
By the boundary conditions shown in Figure 5.2 how many Fresnel reflections are there for an optical ray passing from one fiber into the other? Calculate the increase in fiber transmission if a perfectly index-matched gel filled the gap between the fibers. There are two Fresnel reflections. The first occurs when light passes from the first fiber core into the air gap; the second occurs when light passes from the air gap into the second fiber core. In glass fiber where tolerances of the components are much smaller and fiber ends are truly polished or cleaved, a much smaller increase in transmission occurs when indexmatching a fiber junction (than was observed in this activity). This is not to say that indexmatching is not important in glass fiber. Calculating the increase in fiber transmission due to Fresnel reflections will not show all of the improvements measured in this activity because the gel fills in small imperfections in the plastic fiber core, which reduces losses even further. Often people calculate the increase in transmission simply by multiplying two times the loss, which would be an increase of 7.74%. Although this is approximately right for small losses, the more accurate method of determining the increase in transmission is: 1 =1 (1 R) N
where N is the number of boundaries and R is the loss due to Fresnel reflections
The increase in this case would be 8.21%.
- 12 -
HOMEWORK PROJECT
Fresnel reflection equations have several forms. In reality there are only two fundamental equations. One for each state of optical polarization. These equations are listed below.
sin ( ' ) Rs = ' sin ( + ) tan ( ' ) Rp = ' tan ( + ) Rp is the ratio of
This Fresnel equation can be further reduced if we assume all light rays are normal to the surface. (This is the equation which we used in this lab manual.)
=
( N' N ) 2 ( N' + N ) 2
N' is incident light ray' s material refractive index N is refracted light ray' s material refractive index
In practical applications, many fiber optic light rays are not normal, so the simplest equation which we used here underestimates the loss due to Fresnel reflections. Unfortunately there is not a much better way without defining an equation for the distribution of optical rays and doing some rather tedious Calculus integrals (which might require a computer program to approximate, in any event). It has generally been accepted that the simple equations are accurate enough for most fiber optics work.
- 13 -
600 grit
3 m film
1024
Table 6.2 Calculations for determining losses due to fiber end preparation.
Fiber Termination Phototransistor Current Column #3 Losses %
0.459 0.849
54.1 15.1
- 14 -
Are the increased losses with poorer fiber termination what you expected? Why or why not? Yes. This activity illustrates through measurements that the flatter and smoother the fiber end or tip, the higher the transmission. The data in Table 6.2 show that fiber ends with poor ends or finishes have more losses. Describe how the surface texture caused low phototransistor current with poorer terminations in this activity. Relate your answer to critical angle and fiber end termination. Drawing an illustration may be helpful. The surface texture at the fiber ends directly determines losses when light exits a fiber. The rougher surface texture will cause the light to scatter more, as was shown in Figure 6.1 of the main manual. More scatter has an effect similar to increasing the numerical aperture inside the fiber. Fibers with large numerical apertures require photodetectors with very large active areas to absorb all of the rapidly diverging light rays. In this experiment, when light rays are diverging at a greater rate than is the case with normal operation, some of the rays are not absorbed by the photodetector or, as in this experiment, the phototransistor. Describe how poor fiber termination on both ends, along with light exiting and entering, causes reduced optical power at the photodetector. Poor fiber terminations cause scatter in the incident angle for light rays entering an optical fiber. This scattering is generally uniform and will produce more light rays that are not within the acceptance angle of the fiber core and cladding. These rays, or optical energy, are lost and will not travel down the optical fiber to the photodetector. Poor termination at the photodetector fiber end was described in previous questions. This question is a rewording of a previous question as related to both fiber ends to make the students really think and fully understand what is happening at the optical fiber ends For very short fiber lengths where design margins are good, poor fiber termination is not an issue, but for long-distance telecommunications projects, every bit of optical power must be conserved and utilized. Excess loss anywhere along a fiber system causes system degradation. An example of system degradation might be a very noisy telephone line. For some reason, that line is less than ideal or was poorly designed and has poor signal-to-noise ratio.
- 15 -
HOMEWORK PROJECT
There are many different fiber optic connectors and splice manufacturers. If a student comes up with a company that manufactures fiber optic fusion splicers, we feel that answer counts. The real purpose of the project is to teach students how to find resources. Sources for fiber optic connector and splice companies include: AT & T Microelectronics 3M, Private Network Products Amphenol Corporation Augut Communications Group Ensign-Bickford Optics Company Hewlett-Packard Methode Electronics Norland Products OZ Optics, Ltd. Rifocs Corporation Seastar Instruments Ltd., Optics Division Thomas & Betts Corp., Electronics Division
- 16 -
Table 7.2 Measured values of rise and fall times of four optical sources using the photodiode as the optical detector. LED Infrared Red Green Pinlight Rise time 1.5 s 1.7 s 4.0 s 20.0 ms Fall time 1.5 s 1.7 s 2.0 s 15.0 s
- 17 -
Table 7.3 Calculation of detector upper frequency 3 dB bandwidth using an infrared LED as the optical source. Detector
Phototransistor Photodarlington Photodiode
f3dB
103.0 kHz 17.5 kHz 233.3 kHz
Which photodetector has the largest bandwidth? The photodiode. Table 7.4 Calculations of light source upper frequency 3 dB bandwidth using a photodiode as the optical detector. LED Infrared Red Green Pinlight Rise time 1.5 s 1.7 s 4 s 20.0 ms Fall time 1.5 s 1.7 s 2 s 15 ms
f3dB
233.3 kHz 205.9 kHz 87.5 kHz 17.5 Hz
Which LED is the fastest in Table 7.4 ? Is this what you expected? Why? The infrared LED. Yes. Data sheets for LEDs affirm that infrared types are normally faster than visible ones. Intuitively one might deduce this because, generally speaking, most product specifications supplied by manufacturers are targeted at performance characteristics which buyers/users are most interested in. With infrared LEDs, which are used for communications, we would intuitively expect a higher bandwidth than with visible LEDs. In the case of visible LEDs which are designed for producing the most amount of visible light, it would be an anomaly to find one with fast rise and fall times. Engineers may actually give up speed or bandwidth during the design of a visible device to maximize the visible light output. Why are incandescent bulbs not used as fiber optics light sources? (Use the data in this activity to formulate your answer.) From the data it can be seen that incandescent bulbs are very slow and operate only at very low frequencies. Incandescent bulbs are so slow because current flowing into the bulb heats the bulb filament to a temperature sufficiently high to produce radiation in the visible range. Electrical current into the bulb's filament does not directly produce electrons, unlike
- 18 -
LEDs' response to current. Incandescent bulbs are also not suited for fiber optics because they have short lifetimes and poor linearity. Using information as required from Activity III , determine if a particular LED emits the greatest amount of optical power, has the best optical fiber transmission and the fastest rise/fall times. Which one? If one doesn't meet all the criteria, pick the best LED in each category and list it below. There is no one best LED. Students will need to put on their thinking caps. The answer depends on what performance characteristic the student feels is most important (and which they may include as part of their answer). The infrared LED clearly produces the most phototransistor current as seen in Table 3.1 , and has the fastest rise and fall times. However, the infrared light also has the highest attenuation coefficient of all the LEDs tested (poorest transmission). The red LEDs do not have as high starting optical power, but do have a much lower attenuation coefficient. For applications using long fiber lengths, they may the best choice. The green LED, although an interesting color, offers no advantage in output power, fiber transmission or speed.
HOMEWORK PROJECT
Answers to this question may vary widely. Variables that can affect the student's answers include manufacturer, package type, size, and some devices are optimized for a particular configuration. Ranges that we have found are listed below. You may choose to have students list their sources of information as part of the project.
Device
Photodiode Phototransistor Photodarlington Infrared LED Red LED Green LED
Rise time
1 ns to 3 s 2 to 30 s 40 s to 1.5 ms 1 ns to 1 s 120 ns to 300 ns 450 ns
Fall time
1 ns to 3 s 2 to 30 s 40 s to 1.5 ms 2 ns to 1 s 50 ns to 300 ns 200 ns
- 19 -
Table 8.2 Measured data on the LED drive circuit shown in Figure 8.4. Measurement V ce Vf Rise time Fall time Period 3 dB Data 219 mV 1.97 V 120 ns 100 ns 500 ns
Table 8.3 Measured data for the highspeed LED drive circuit shown in Figure 8.5 . Measurement V c (LED on) V c (LED off) Rise time Fall time Period 3dB Data 3.36 V 1.56 V 20 ns* 4 ns* 50 ns* - 20 * 40 MHz oscilloscope may not be able to measure these fast transition times, but that is revealed as the circuit capability when using a 350 MHz oscilloscope as a measuring instrument.
Table 8.4 Measured data for the analog LED drive circuit shown in Figure 8.6. Measurement Ic V c (.5-volt input) V c (distorted) V i (distorted) Period 3 dB Data 7.9 mA 3 V p-p 7.8 Vp-p 1.4 Vp-p 128 ns
Using the measured rise time from Table 8.1 , calculate the 3 dB bandwidth for the circuit shown in Figure 8.3.
f3dB = f3dB .35 240 10 9 = 1.45 MHz
How does the calculated bandwidth compare to the measured bandwidth? The measured bandwidth can be calculated from the following equation. (It is assumed that students will have knowledge of this general equation from their electronics experience.) MEASURED f3dB = 1 670 ns = 1.49 MHz - 21 -
The calculated 3 dB bandwidth from the rise time was determined to be 1.45 MHz and measure was 1.49 MHz. It would not be uncommon for the students' measured 3 dB bandwidth and calculated bandwidth from risetime to be 10% apart from each other. Calculate the average current used by the LED driver in Figure 8.3, assuming it is being driven at a 50 percent duty cycle.
Iavg = I PEAK Duty cycle + I MIN (1 Duty cycle) Iavg = 19.2 mA .5 + 0 .5 Iavg = 9.6 mA
Using the measured value for Vf , calculate the current through the LED in Figure 8.4.
What is the average current though the circuit shown in Figure 8.4, assuming that it is being driven at a 50 % duty cycle?
Iavg = Iavg 31.9 mA + 20.2 mA 2 = 26.1 mA
Comparing the peak current and average current for the circuits in Figures 8.3 and 8.4 , which would cause the greatest power supply ripple? By how much? To compare power supply ripple we shall use the ratio of the peak current divided by the average current. Shown below is the ripple factor for the circuits in Figure 8.3 and 8.4 .
rfFigure
8.3
19 mA 9.6 mA =2 =
rfFigure
8.4
31.9 mA 25.6 mA
= 1.25
The circuit in Figure 8.3 has a much higher power supply ripple, as can be seen by the equation which shows the Ipeak to Iaverage ratio to be 2, as compared to the same ratio for Figure 8.4, which is 1.25.
- 22 -
What is the difference between the circuits in Figure 8.3 and 8.4 ? (HINT: Consider inverted and non-inverted functions.) In Figure 8.3, the LED operation follows the drive signal state, i.e., driver signal high = LED on; drive signal low = LED off. In Figure 8.4 the LED operation is inverted from the drive signal state. Also, the circuit of Figure 8.3 draws supply current only when the drive signal is high. The circuit of Figure 8.4 always draws current, either through the LED or transistor. Assuming hfe is 100 for the PN2222, calculate the DC LED current drawn for the circuit shown in Figure 8.6 with no input signal.
Ic =
( 560
47 +
10 V ) .7 V
( 560 + 4700 )
100
= 7.0 mA
What is the maximum linear voltage swing of the circuit shown in Figure 8.6? (HINT: Determine the answer from empirical data.) The value of the output voltage just below where signal distortion occurs, or slightly less than 7.8 volts peak-to-peak. What is the 3 dB bandwidth of the circuit shown in Figure 8.6?
Measured f3dB = = 1 Period3dB
HOMEWORK PROJECT
As is typical with design problems there is often more than one solution. Here is one possible solution, shown with different transistor part numbers suitable for this circuit.
+ 8 volts
.1 f
56
Pulse Gen.
1, 3, 5, 9, 11, 13
470
1025
- 23 -
Table 9.2 Measured data for various termination resistors, Rf , in the circuit shown in Figure 9.4. Rf 100 k 47 k 10 k 10 k ||.001 f Vp-p 4.5 V 3.1 V 0.66 V 0.66 V tr 7.2 s 4.2 s 1.6 s 1.5 s tf 7.4 s 5.0 s 1.1 s 1.1 s
Table 9.3 Measured data for various termination resistors, Rf , in the circuit shown in Figure 9.5. Rf 100 k 47 k 10 k 10 k ||.001 f Vp-p 2.5 V 1.9 V 0.6 V 0.6 V tr 1.6 s 1.3 s 1.0 s 1.0 s tf 1.1 s 1.5 s 1.2 s 1.2 s
- 24 -
Table 9.4 Measured data for various termination points in the circuit shown in Figure 9.6 (a). Location V emitter V 4069 Vp-p 4.8 V 5.0 V tr 4.0 s 0.5 s tf 7.0 s 0.4 s
Table 9.5 Measured data for various termination resistors, Rf , in the circuit shown in Figure 9.6 (a). Resistance Phototransistor Photodarlington Value 213 22
Calculate the bandwidth of the receiver shown in Figure 9.4 with the 10 k gain resistor installed using the equation: f3dB = .35 1.6 10 6 = 219 kHz
Calculate the bandwidth of the bipolar transistor receiver in Figure 9.5 using the 10 k feedback resistor installed. f3dB = .35 1.0 10 6 = 350 kHz
How much more gain does the photodarlington have than the phototransistor in the circuit in Figure 9.6 (a) ? (Hint: Use the ratio of the resistors that were in parallel with the 390 resistor which remove periodic signal at the output of 4069.) With a given optical input, the photodarlington requires a smaller load resistance than the phototransistor to develop the same output voltage. The relative gain can be found by determining the ratio of the resistors that results in the same signal level for each device:
GRe l . = =
Rl ( phototransistor ) Rl ( photodarlington )
213 22 = 9.7
The photodarlington in this activity had a gain 9.7 times greater than the phototransistor.
HOMEWORK PROJECT
Students' written paragraphs must be interpreted by the instructor. A quick overview of many fiber optic companies is found in the Photonics Corporate Guide, Volume I and the Fiberoptic Product News Technology Reference. You may find it useful to purchase these publications. To order, see page 53 of the main manual for more information.
- 26 -
IF-LM-A